“Despite its current misadventure with Linux, Sun isn’t in the generic desktop computer business. The Java desktop is cool, but it’s a solution driven by necessity, not excellence. In comparison, putting Mac OS X on the Sunray desktop would be an insanely great solution for Sun.” read the rest macnewsworld.com
I believe Apple has a great partner IBM who has the facilities to deliver a great processor. With increasing sales of G5’s now that the new iMac is out. It’s good for all sides, Apple and IBM to futher develope beyong the PPC including the Cell.
Isn’t Sun the world’s largest Linux vendor, now, due to gigantic contract in Asia? Yes, I said world’s largest.
Oh yes, the supposed 200 million copies of JDS. I wonder why there hasent been any talk about the actual deployment. Maybe because it was just BS?
The combination makes total sense to me, last year I administered Sun servers with an Powermac and it is a perfect marriage.
It really interests me how these new processors can eat away from the Intel/AMD market.
When the demand gets at peak levels who will Big Blue serve first?
A lot of people write that Apple should convert their OS to intel architecture, but maybe that never gonna happen but instead Microsoft converts Windows to Cell/ Power processor.
/happy newyear/
I don’t really think this guy is credible. For example, he says he doubts that IBM really had trouble with their 90nm process, and uses the production of Cell at 65nm as evidence. Well, the 65nm Cell processors won’t use IBM’s process — they’ll use Sony’s 65nm process at their Nagasaki plant. He also makes some dubious claims about the G5’s simpler ISA causing it to be simpler to manufacture, even though the G5’s internal architecture is every bit as complex as a P4’s or Athlon’s.
His P615/Xserve comparison is also totally off-base. The two machines aren’t even in the same league. The P615 is a Power4+ with 1.5MB of L2 and 8MB of L3. The G5 in the Xserve as 512KB of L2 and no L3. The P615 has 6 PCI-X slots, the G5 has only 2. The P615 has all sorts of reliability features like ChipKill, etc, that the Xserve doesn’t have. It’s literally Apples and oranges.
In any case, the article is years too late. Apple bet on IBM saving PPC, and that bet is paying off. PowerPC is taking off right now, and SPARC is falling futher behind in every metric. An Apple -> Sun connection would make absolutely zero sense right now.
Yeah I read the article, wondering if the guy even knew what end was up. Then when he stated running MacOS on Linux Servers, was kinda like saying, I am totally clueless and don’t know the difference between OS and Apps.
Plus, people who keep up with processers know, cell has a altec clone that with tweaking can be made to act like a atlec chip. There is a lot of things point to the guy drank to much egg nog for christmas.
Maybe Apple should become CPU agnostic. Just sell an OS and a killer set of apps, and leave the drudgery of hardware manufacture to companies better equipped to handle it. After all, what’s the margins in the hardware market anyway? In the first quarter of ’04, HP earned $45m on $6bn of PC sales. It’s tiny! Now, Apple’s margin is a lot higher, but only because they sell a niche product. If they wanted to become mainstream, they couldn’t maintain that sort of profit margin. Meanwhile, the profit margin in the software market is in the double-digits.
What’s “altec”? Do you mean AltiVec?
Per the article (assuming you actually read to the end): “This story was originally published on July 8, 2004, and is brought to you today as part of our Best of ECT News series.”
Nice of OS News to publish something almost 6 months old.
Yeah.. BS… heh… “I don’t like it. So I will just pretend it didn’t happen and call it BS.” Typical anti-Sun zealot response.
Well, newsflash. The 200 million copies is not happening overnight. It was designed to be an incremental migration. And yes, it is in progress.
And as far as OS X on Sun Ray… Not gonna happen. Sun Ray’s are overpriced and underpowered, just like Apple’s machines. That and Apple would have to do a better jon of keeping OS X up to date with the latest releases of Java.
Apple + Sun? No way. Sun isn’t even interested in OpenOffice/StarOffice on Mac.
And Apple should port OSX to SPARC while Sun is moving to Opteron? What should happen to iBooks and PowerBooks? Replace them with SPARC notebooks that have a battery life of only ~2 hours? Yeah right…
Well, newsflash. The 200 million copies is not happening overnight. It was designed to be an incremental migration. And yes, it is in progress.
And you know how? A source with a project update? Do you have anything to backup your claim? Still waiting…
Yeah, right.
I’d be the first one in line to buy OSX for x86, but until the Apple board boots Jobs out (again), that’s not going to happen. Jobs’ hardware fetish is probably still in the driver’s seat up in Cupertino.
I agree with your assessment that it’s ridiculous for Apple to go Sparc. Does a Sparc consumer processor setup have anything over G5s, AMD64, or intel x86 now or in the near future?
I did like how he talked about the multicore future though and it could be IBM kindof hanging Apple out to dry without major rewrites of OSX/apps and how Linux plays into this. Actually, if DragonFLY BSD is designed specifically for the multicore future then we might see it as a player. I don’t know enough kernel development to know how much of the linux core kernel will have to re-wacked to take advantage of multicore. How much different is it than regular SMP? Schedulers? VM?
Apple is a HARDWARE company, that makes great software on the side.
Apple rely’s on a tight hardware software intergration, Something DELL, HP, Levono, etc can NEVER achieve with MSFT as the ruling desktop OS.
Dell can more easily become like Apple and boost margins, simply by working with a linux distro to closely tie hardware and software.
I see the future of computing where companies don’t try to outdo each other with just software but by controling the entire user experience. Something that CAN’T happen with MSFT as it currently is.
Viruses will have a tough time because each core OS will be written for a different architecture. You can limit damage and enhance your overall performace, by merging different style chips in modular clusters. Certian chips perform different functions faster, break the tasks up to the chips that will best perform those tasks.
If you search Google (http://www.google.com/search?q=site:www.sun.com+china+jds), you’ll see hints here and there it’s still going on. One was from a chat discussion with Scott McNealy on November 15, 2004. Another tidbit was from June.
“Sun Ray’s are overpriced and underpowered…”
How so? They are zero administration remote frame buffers for the server. Their incremental cost is practically nothing from a big picture point of view and they are as fast as the server you deploy. Imagine have a slice of a 12 or 24 CPU server with 100GB of RAM and FibreChannel RAID.
Also, Sun is completely honest and forward in their marketing materials that Sun Ray is not appropriate for 3D work, such as CAD. Get a workstation for that. Sun Ray is very appropriate for everyone else, however (help desk, clerical, administrative, etc.).
and suddenly we are right back in the early 80’s. sure, stuff like java and .net helps a bit but haveing to toss the whole computer every time one feel like upgradeing is a bit overkill…
that is unless the computer is basicly the cpu, a video board and not much else. lets put the users apps and desktop enviroment on a external hd or something (usb2 and/or firewire). the kernel and drivers are put on a flash card in the comp itself. after boot the kernel fires up a very minimal gui that searches any external storage units looking for user profiles and app folders. every user profile found is presented in a list and the user can then log in.
and how about the storage media can double as a music and video player? so that on its own it can play back audio useing headpones or a stereo hookup, and video useing svideo or similar. hell, have it show images over the video connection to
apps should be wrapped in selfcontained archives with no need to install or uninstall. want to get rid of a app? delete said archive.
hmm, maybe one could make a laptop where the storage media could dock so that you didnt have to worry about cables and so on. just bring the whole thing along. come home, pop the storage out, hook it to your desktop and continue working.
It makes no sense (at this stage in the game) to port OSX to x86. I think 2004 represented the very beginning of the end for the ancient ISA. If Apple was to port OSX to x86, they would have (if not should have) begun working on that right after the original PPC launch.
Apple has a range of technologies that are widely deployable across much of the market, but has yet to leverage their position in this respect. Spreading the gospel of OSX from the embedded client to the data center is possible only by embracing commodity hardware. The only way that Apple can fight the onslaught of commodity software systems (read: mostly Linux) is by reacting more quickly to coming revolutions in computer architecture. Forget x86 and SPARC (and their 64-bit incarnations); the future lies in Apple’s ability to fully leverage the IBM Cell processor, because MS won’t be quick to do so.
I think 2004 represented the very beginning of the end for the ancient ISA
Not even close. The x86 ISA is here for years and years to come. But it’s all RISC under the hood anyway. AMD64 still has all the x86 instructions.
If Apple was to port OSX to x86, they would have (if not should have) begun working on that right after the original PPC launch.
How much assembly language do they have to port? It’s Mach and BSD. Do they have C wrappers for some of their altivec stuff?
Forget x86 and SPARC (and their 64-bit incarnations); the future lies in Apple’s ability to fully leverage the IBM Cell processor, because MS won’t be quick to do so.
You’re a fool if you discount Intel and AMD
Apple is a HARDWARE company
I know that. My point is that the hardware market isn’t a great one. If you ship in volume like HP and Dell, your profit margin must be tiny in order to compete. If you want to have a high profit margin, like Apple does, then you have to price yourself into niche, like Apple has. It’s like the high-end car market — it’s a tiny fraction of the overall computer market and always will be. If Apple wants to get bigger, they have to do something else. Since they can’t compete with HP and Dell in the volume market, the only thing I can see for them is software.
There is no reason they have to give up their hardware business to do this. People who would otherwise buy a Mac aren’t going to avoid doing so just because they could get OS X on a PC. Meanwhile, they could make quite a nice chunk of change selling $129 OS X upgrades to PC users once every 9 months.
So I suppose IBM can’t work on their POWER line of processors and cell at the same time? Not to mention the shame and horror of not hitting 3GHz yet. I mean I don’t know about everyone else here but two 2.5GHz G5s just can’t suffice it for me these days. We’d better switch over to an entirely new platform and pronto. It’s not like we have anything invested in the PPC architecture. Well aside from all the commercial software.
Sense we are on speculatory mode. Why not have AMD, IBM, Sun, and Apple coumbine their chip making prowless to counteract the market dominancy of Intel?
Makes sense to me, though of course many issues would have to be overcame.
What was that…IBM & Sun Microsystems working together on a chip. Go and have a read of Jonathan Schwartz’s Blog ( http://blogs.sun.com/roller/page/jonathan ) then come back and make some informed comments. It make’s sense to me too if I try not to think too hard.
Apple has taken a good step by establishing a partnership with Big Blue (IBM)…the G5 CPU is impressive, and the right chip. Instead of APple going for a different CPU they should fully utilize the G5 first. You see the G5 is a 64bit CPU whereas OsX is a 32bit Os. Clearly Apple is behind in exploting the G5 advantage. A G5 machine is currently an over priced adn under-utilized CPU, at least until Apple makes Osx truly 64bit (it is merely 32bit now). I for one hope that Apple does that soon, but for now, the revenue is clearly on the iPods, and APple should concentrate on that, put the Macs on the backburner, get the iPods, and then either transition to fully 64bit or dump the desktop. No room for sentimentalism in bussiness, iPods are the tru shiny stars in Apple future right now.
Go Apple go!
PowerPC is taking off right now, and SPARC is falling futher behind in every metric. An Apple -> Sun connection would make absolutely zero sense right now.
How can you say that with Fujitsu’s SPARC64 topping 2GHz and the upcoming release of Niagra? SPARC is certainly outdoing POWER for the time being…
To port OSX to x86 would make as much sense as porting a B&O DVD-player firmware to a SAMSUNG DVD-player. None at all.
Q: What is Apple’s mission statement?
A: Apple ignited the personal computer revolution in the 1970s with the Apple II and reinvented the personal computer in the 1980s with the Macintosh. Today, Apple continues to lead the industry in innovation with its award-winning desktop and notebook computers, OS X operating system, and iLife and professional applications. Apple is also spearheading the digital music revolution with its iPod portable music players and iTunes online music store.
Notice how they use words as inovation and revolution. It is not part of their mission to dominate the market. They simply wants to be the best.
If they want to continue to be the best, they would need some deep pockets, and for that they need larger market share and profits.
It is not part of their mission to dominate the market. They simply wants to be the best.
Don’t be ridiculous. Apple is a publicly traded company. The only thing they care about is profit, both as a matter of practicality and one of legal obligation to its shareholders.
Ofcourse they only care about profit. I’m saying that you don’t have to have a large market share to have large profits. Take the company I work for as example we have about 2% of the market, and one of our main goals is to be the most expensive, because being most expensive gives our customers the confindence that they get the highest quality. We have no plans to lower our quality just for growth.
I should add that this strategy probably only works as long as the market it self is growing.
The Power5 at 1.9Ghz is about the same as the 1.89GHz SPARC64V on SPECint, and 50% faster on SPECfp. Meanwhile, the much maligned Itanium2 is 20% faster on SPECint and 50% faster on SPECfp. Out of all the big competitors (Opteron, Xeon, POWER5, Itanium2), is dead last in both SPEC benchmarks. The SPARC is just not an interesting architecture at this point — the performance crown has been tossed back and forth between IBM, Intel, and AMD for awhile now (Itanium is great at FP, Athlons/Opterons are great at integer, and POWER5’s are great at both), and SPARC has consistently been dragging up the rear.
Now, the CPU isn’t everything — system architecture counts for a lot. But when you’ve got a brick of a CPU like most Sun systems, there is only so much you can do to the system architecture to make it perform well.
How can you say that with Fujitsu’s SPARC64 topping 2GHz and the upcoming release of Niagra? SPARC is certainly outdoing POWER for the time being…
I’m not impressed with Sun hardware. We just purchased a new Sun server, dual proc with 2 GB of ram w/ Solaris 9 on it. The thing is slow. I develop java apps on my PC before I upload it to our app server, and a 1.7Ghz P4 with 512MB is twice as fast at executing the sql queries and serving the page.
If our IT department spends over 2,000 on a server, I expect it to be faster than my 2.5 year old Dell PC. I know it’s only running the UltraSparc III(i?), but you would think that a brand new server would perform just a little better than it does.
Just how much would I have to pay to get decent hardware from sun? 10k?
Ofcourse they only care about profit.
No, you said, and I quote, “They simply wants to be the best.” Being “the best” isn’t necessarily a good way to make a profit. Purchasing decisions are a cost/benefit analysis. That means that adequate and cheap is usually easier to sell than superlative and pricy.
Given that the PC market has steady growth, the only way for Apple to keep increasing it’s profits, aside from side businesses like the iPod, is to increase it’s share of the market. There are two ways to do this: release simpler products (eg: nobody needs GigE in a corporate econo-box), and compete in the volume market with Dell and HP, or focus on software and let others take care of selling low-end hardware. Since Apple cannot possibly compete with Dell and HP in the volume market, they can either go the software route, or remain in their current high-quality high-price niche market.
One year to almost the day I wanted a powerful notebook(not a portable). What do I hear from Apple. NADA!. I get top-o-line 1.2 g4. Give me a break. I bought a 3.2 ghz p4 with a gig of ram that I run windows and linux on (at the same time http://ww.colinux.org) and apple is telling me that a freaking 1.4 ghz is all i need.
And it was a grand less than top-of-line.
Give us power users (programmers/developers) a reason to choose u.
Actually, Jobs is such a fucking idiot that once the NeXT crowd bogarted in, Dylan development hit the skids.
It’s not as simple as just slapping it onto x86 hardware. There are major support issues for them if the decided to port OS X. Think of all the drivers that would suddenly need to be supported. How many designers using Mac do you know who would be able to handle the problems that would arise with configuring their system ? Mac flaunts the fact that they’re computers are easy to use and they would end up disapointing a lot of people. Besides, it’s true that hardware isn’t the most profitable market at the moment, but next time you’re in a PC shop go compare the prices for similar performance of Macs vs PCs. Then tell me wether they aren’t doing pretty well. A lot of people are just buying the cool looking hardware. Hell even I want to !
blah, blah, blah. Fuck it man. Jobs has a hardware fetish and that’s all there is to it.
Fine, they think they’ve got their margins, but that’s their business.
APPLE IS NOT A HARDWARE COMPANY.
Get over it. They don’t do processors. They don’t do components.
They make cases!. That’s it.
Give us OSX on a processor that we choose before Apple slides into oblivion.
Give us OSX on a processor that we choose before Apple slides into oblivion.
Couldn’t agree more.Apple is a good assembly company and OS designer.
Give us OSX on a processor that we choose before Apple slides into oblivion.
Give me Win2k3 on Power arch then.
That would spell death for Apple, they are a hardware company first, not second. Selling hardware with their own software allows for a tighter intergration and a better overall experiance for the customer. I don’t see that happening if OS X made it’s way to X86.
Apple could move to x86 and only officially support installations on a limited selection of the best hardware available.
If the user wants to run on the cheapest stuff they can find they shouldn’t expect Apple to provide drivers for it.
Apple could still create thier own Macintosh hardware and provide full support for it.
Of course I’m waiting for the $500/$600 headless Mac, if it is everything that has been reported I’m going to have to start making room in my budget for it.
I don’t think anyone is claiming that OS X on x86 would be as “tight” an experience as OS X on PPC. People who want that “tight” experience would continue to buy OS X on Apple hardware. The point is that OS X on PPC is a premium product, like a Mercedes or BMW. Very few people actually buy premium product. If you just concentrate on those, you are limiting yourself to a few percent of the market. After all, it’s interesting to note that Daimler now sells both S500’s and Dodge Neons, and BMW now sells both M5s and Volkswagen Jettas.
The $499 McMac, the computer Happy Meal of the future!
Could be. We agree with the approach. Tied into an iPod, while leveraging AltiVec and the G4. With the G5 still evolving could be an excellent strategy for Apple. It is an upsell and it might work. In the Apple’s fourth quarter, 23 percent of its $2.35 billion in revenue came from iPod sales.
Why not stick with the G4? Could it be a nice fat client for Sun servers? Freescale just has silicon to offer and most Apple computers sold use it. Apple may have gone in the other direction and have alot more working for them than the missing paddle…;-)
R&B
“I develop java apps on my PC before I upload it to our app server, and a 1.7Ghz P4 with 512MB is twice as fast at executing the sql queries and serving the page.”
One of the flaws I noticed about many PC Java developers is that they think trivial benchmarks tell them everything they need to know about a platform. “Oooh, this system is slower at X, so let’s trash it entirely and go to this other system which is cheaper up front and is on the cover of JavaPro!” It pretty much burns them, because they don’t know all the variables, but they never realize it until they are laid off (this takes years, because their managers are often from the same breeding pool and it takes a few years to run a company into the ground).
Sun still manages to break records on a number of application benchmarks, even with measly systems like your UltraSPARC IIIi one. There’s more to life than just an SQL query. What are the bottlenecks in your system? Have you done profiling? What about under load? What happens when you throw tens of thousands of queries against it? Does your PC look as good, then, with its IDE disks and crappy OS?
…nothing to see here, but yet another crank trumpeting the latest Apple death knell. These days they seem to revolve around, OMG WTF I own Intel/AMD stock so Apple should use x86, or the even further out partnership with Sun?!
One of the flaws I noticed about many PC Java developers is that they think trivial benchmarks tell them everything they need to know about a platform.
Heh you’ve a bit of a cheek to talk about how bad it is he generalises based on one silly variable after that..
What are the bottlenecks in your system?
As far as I can tell, just the processor is the bottleneck (perhaps the OS?). It has a nice SCSI Raid, and has a fibre channel connection to our storage devices.
Have you done profiling?
I haven’t personally done profiling, that is what we pay our administrators for. She tells me that it is as good as it’s going to get.
What about under load? What happens when you throw tens of thousands of queries against it?
It goes slower. We run Informix 9.4, which does a pretty good job. We never really have a huge load. Probably at most we have 20 people accessing the server at one time. However, when it takes 10 seconds to execute the search query on the Slowaris box and 0.700 seconds on the Windows box (my workstation, not a server) with the crappy disk and OS, something is obviously not right.
I’ve tried arguing with the DBA and the Solaris Administrator, however, they tell me that I just have to live with it (and I do).
In my personal experience, Sun doesn’t provide a very good price/performance ratio with their SPARC technology (which is probably why they are beginning to offer more Operteron offerings).
I do love my SunRay though, and I can’t wait until sun starts bundling a version of Gnome that is newer than version 2
.
Linux SunRayPPC? Nice!
Are you comparing a development PC to the server setup?
“As far as I can tell, just the processor is the bottleneck (perhaps the OS?).”
Solaris scales really really well. It sacrifies a little single-threaded performance in favor of near-linear scaling on SMP and under load.
“I haven’t personally done profiling, that is what we pay our administrators for. She tells me that it is as good as it’s going to get.”
Using profiling tools is the developer’s responsibility. Should the administrator be expected to care about which functions in your program should be optimized? In the case of profiling the SQL query itself, it is useful to get the help of the DBA, though.
“Probably at most we have 20 people accessing the server at one time. However, when it takes 10 seconds to execute the search query on the Slowaris box and 0.700 seconds on the Windows box…”
A 1400% difference in query speed? This performance difference should be a red flag to investigate your configurations from the OS on up, because there is something very wrong with how your servers are set up. There must be some caching going on that you don’t know about on your PC.
After all, it’s interesting to note that Daimler now sells both S500’s and Dodge Neons, and BMW now sells both M5s and Volkswagen Jettas.
Huh? When did that happen? Last time I looked Audi was owned by Volkswagen, not BMW.
Duh
You’re right, of course.
Apple isn’t a hardware company, nor are they a software company. They will dip their toes in both when necessary and come up with something. But what Apple really is, is a presentation company. They sell an image. Their products are as much about style as they are about capability. Look at their MacOS. Look at their I-pod. Look at their Apple Store. They are all about presenting a fully finished product, and will do anything to get that done.
This article came from a mac news magazine. why are people complaining this guy is a Apple hater/basher? Why would a Macintosh magazine article bash the very product it’s writing about?